moderndid.wboot_drdid_ipt_rc1#
- moderndid.wboot_drdid_ipt_rc1(y, post, d, x, i_weights, n_bootstrap=1000, trim_level=0.995, random_state=None)[source]#
Compute IPT bootstrap estimates for control-only doubly-robust DiD with repeated cross-sections.
Implements the bootstrap inference for the doubly-robust difference-in-differences estimator using inverse probability tilting (IPT) for propensity score estimation. This version uses outcome regression on control units only.
- Parameters:
- y
numpy.ndarray A 1D array representing the outcome variable for each unit.
- post
numpy.ndarray A 1D array representing the post-treatment period indicator (1 for post, 0 for pre) for each unit.
- d
numpy.ndarray A 1D array representing the treatment indicator (1 for treated, 0 for control) for each unit.
- x
numpy.ndarray A 2D array of covariates (including intercept if desired) with shape (n_units, n_features). The first column MUST be an intercept.
- i_weights
numpy.ndarray A 1D array of individual observation weights for each unit.
- n_bootstrap
int Number of bootstrap iterations. Default is 1000.
- trim_level
float Maximum propensity score value for control units to avoid extreme weights. Propensity scores for control units greater than or equal to this level will lead to the observation’s weight being set to zero. Default is 0.995.
- random_state
int,RandomStateinstanceorNone Controls the random number generation for reproducibility.
- y
- Returns:
numpy.ndarrayA 1D array of bootstrap ATT estimates with length n_bootstrap.